FR2835667A1 - Monotonic-phase filter control method, e.g. for baseband filter in direct conversion receiver, using configuration stage in oscillator mode for setting cut-off frequency - Google Patents

Abstract

The filter is set up by firstly operating in oscillator mode, during which time the oscillator frequency is measured. The characteristics of the filter are adjusted in order to match the required cut-off frequency for the filter, as closely as possible, within a determined degree of tolerance. The filter control procedure includes a phase in which the filtering circuit (FF10) operates as an oscillator, and the frequency of operation of the oscillator is determined. The characteristics of the device are then corrected, taking account of the measured frequency of oscillation and a predetermined relationship between the frequency of oscillation and the theoretical cut-off frequency of the filter. This gives the circuit a cut-off frequency which is equal to the frequency of oscillation, within a close tolerance. After this adjustment phase, the filter moves into a working phase, in which the filter carries out its filtering function. When the filter has an order greater than three, the filter operates in oscillator mode by re-looping to itself with phase inversion. When the filter operates with order less than three, an additional series filter is provided to form the oscillator loop.

Description

<Desc / Clms Page number 1>

 Method for adjusting the cut-off frequency of an electronic filtering system, and corresponding system.

 The invention relates to electronic filtering systems, and more particularly their calibration, that is to say the adjustment of their cut-off frequency.

 The invention is advantageously but not limited to wireless communication systems, and more particularly to cellular mobile telephones, the reception and transmission chain of which incorporates filtering devices.

 In the reception circuits of cellular mobile telephones, after conversion of the high frequency analog signal into a low frequency signal, it is imperative to filter the latter so as to keep only the useful information.

 In direct conversion receivers, this filtering is carried out by low-pass filters. However, it is imperative that the cutoff frequency is known with good accuracy. And, in integrated systems, this cut-off frequency can vary up to 30%, such a variation being linked in particular to the manufacturing process and the operating temperature.

 It can then result, during the operation of the telephone, a loss of the useful signal if the cutoff frequency decreases too strongly, or else a degradation of the useful signal due to a bad rejection (rejection) of the jammers if the cutoff frequency increases too much strongly.

 This is the reason why it is necessary to calibrate the filter, i.e. to adjust the cutoff frequency of the filter to a known value, in this case the theoretical cutoff frequency, with greater precision. .

 Currently, to perform this calibration, it is possible to use a phase-locked loop, the oscillator of which is produced with elements similar to those of the filter to be calibrated, in particular a resistive-capacitive network. The filter calibration is then carried out in

<Desc / Clms Page number 2>

 calibrating the oscillator, and applying the same correction on the oscillator and on the filter, for example by switching capacitors of the resistive-capacitive network.

 However, there is always a matching error between the external phase locked loop and the filter to be calibrated.

 Consequently, the correction determined on the phase locked loop external to the filter, and applied to the filter itself, is not exact because of this pairing error.

 Furthermore, the presence of this matching error leaves very little margin for imprecision in the technological implementation of the rest of the integrated circuit, which is particularly disadvantageous when the cut-off frequency must be adjusted with a low tolerance, slightly greater than the matching error.

 Finally, the presence of an external phase locked loop for calibration increases the surface area of the integrated circuit, which has an impact on the cost of production and the size.

 The invention aims to provide a solution to these problems.

 An object of the invention is to propose a filtering system the calibration of which is not marred by a matching error.

 The invention also aims to propose a filtering system whose calibration constraint requires only a very small increase in the surface area of silicon.

 The invention therefore proposes a method for controlling the operation of a filtering device with monotonic phase, having a theoretical cut-off frequency. A monotonic phase filtering device is for example a high-pass filter or a low-pass filter.

 According to the invention, this control method comprises a calibration phase in which the filtering device is operated as an oscillator, the oscillation frequency of the filtering device is determined, and the characteristics of the filtering device are corrected taking account of the determined oscillation frequency and a pre-established relationship between the oscillation frequency and the cutoff frequency

<Desc / Clms Page number 3>

 theoretical, so as to give the filtering device a cutoff frequency equal to the theoretical cutoff frequency to within a tolerance.

 Furthermore, the control method includes a working phase in which the filtering device performs its filtering function.

 In other words, the invention provides for using the filtering device itself as an oscillator. No external device is therefore used for the calibration, and therefore any pairing error between such an external device and the filtering device itself is eliminated.

 When the filtering device has an order greater than or equal to 3, the filtering device is operated as an oscillator by looping it back on itself with a phase inversion. Indeed, when the order is greater than or equal to 3, the evolution of the phase of the filter as a function of the frequency is such that there is a frequency value for which the phase of the filter is equal to +180 or-180 , which allows its oscillation.

 On the other hand, when the filtering device has an order less than 3, for example an order 2, the value of 180 or de-180 can never be reached for any frequency value whatsoever. Consequently, to allow the oscillation of the filtering device, and consequently its calibration, an identical auxiliary filtering device is then advantageously connected in series to the filtering device. And, the assembly formed of the two filtering devices is operated as an oscillator by looping this assembly back on itself with phase inversion. This allows the filtering device to be calibrated.

 When the input of the filtering device is a virtual ground, that is to say when the filtering device is for example formed of an operational amplifier, the loopback is advantageously controlled with a current injected onto the virtual ground of the device filtering, so as to adjust the amplitude of the oscillations. So, we

<Desc / Clms Page number 4>

 advantageously choose the value of the current so that the filter works in the linear operating zone.

 When the cutoff frequency of the filtering device is defined by a resistive-capacitive network, the characteristics of the filtering device are advantageously corrected by modifying the capacitive value of the resistive-capacitive network.

 The invention also relates to an electronic filtering system, comprising - a monotonous phase filtering device having a theoretical cut-off frequency, - first controllable means connected to the filtering device and capable of being activated or deactivated in response to a control signal, so as to operate the filtering device as an oscillator when they are activated, the filtering device performing its filtering function when the first means are deactivated, measuring means capable of determining the oscillation frequency of the filtering device, correction means capable of correcting the characteristics of the filtering device taking into account the determined oscillation frequency and a predetermined relationship between the oscillation frequency and the theoretical cut-off frequency, so as to give at the filtering device a cut-off frequency equal to the tea cut-off frequency orique to within tolerance, and control means capable of delivering the control signal.

 According to an embodiment in which the filtering device has an order greater than or equal to 3, the first means include a loopback stage with phase inversion connected between the output and the input of the filtering device.

 According to an embodiment in which the filtering device has an order of less than 3, the first means include a loopback stage with phase inversion connected between the output and

<Desc / Clms Page number 5>

 the input of the filtering device, this looping stage comprising an auxiliary filtering device identical to said filtering device, the two filtering devices being connected in series.

 According to one embodiment of the invention, the input of the filtering device is a virtual mass and the looping stage comprises means capable of controlling the looping with a current so as to adjust the amplitude of the oscillations.

 According to one embodiment, the filtering device has a differential structure and comprises a differential operational amplifier, and the control means comprise a differential pair of transistors, for example field effect with isolated gates, the respective sources or transmitters of which are connected, whose respective grids or bases are connected to the two outputs of the differential amplifier, and whose respective drains or collectors are connected to the two inputs of the differential amplifier, and a current source capable of being activated or deactivated by response to the control signal, and connected to the sources of the two transistors.

 The invention also relates to a remote terminal of a wireless communication system, for example a cellular mobile telephone, incorporating a filtering system as defined above.

 Other advantages and characteristics of the invention will appear on examination of the detailed description of embodiments and implementation, in no way limiting, and of the appended drawings, in which: FIG. 1 illustrates, schematically, a reception chain of a cellular mobile telephone incorporating a filtering system according to the invention; - Figure 2 illustrates, in more detail but still schematically, an embodiment of a filter system according to the invention; and FIG. 3 illustrates, in more detail but still schematically, part of the filter system of FIG. 2.

 In FIG. 1, the reference TP designates a cellular mobile telephone comprising, in a conventional manner, an ANT antenna followed

<Desc / Clms Page number 6>

 a reception chain comprising a low noise LNA amplifier at the head. Downstream of the LNA amplifier, two mixers MIX1, MIX2 are connected, respectively receiving two local oscillator signals, mutually phase-shifted by 90, and respectively referenced OSC1 and OSC2. These two mixers transpose the signal from the LNA amplifier into baseband.

This is called a direct conversion or zero intermediate frequency receiver. The two processing channels are then in phase quadrature and are respectively called by the skilled person channel I and channel Q.

 Conventionally, each processing channel comprises an amplifier with controlled gain AGC1, AGC2, framed by low-pass filters FF10 and FF11 on the one hand, and FF20 and FF21, on the other hand.

 The analog reception chain ends with an analog-digital ADC conversion stage which forms the input of a digital processing block BN, which in particular comprises processing means such as a processor PR, performing in particular demodulation as well as conventional channel decoding treatments.

 We will now describe, with particular reference to Figures 2 and following, one of these filtering systems, in this case the filtering system FF10, it being understood that the other filtering systems have an identical structure.

 In FIG. 2, a differential structure is shown, it being understood that the invention also applies to a single-entry structure.

 The filtering system FF10, advantageously produced in an integrated manner on a silicon chip, comprises a filtering device proper, here formed of an active filter, comprising a differential operational amplifier AOP associated in conventional manner with a resistive-capacitive network RC to define its cutoff frequency.

 The filtering device is looped back on itself by an ETRB looping stage, here comprising a transducer formed by a differential pair of field effect transistors with isolated gates

<Desc / Clms Page number 7>

 (MOS transistors) for example with N channel, and respectively referenced MN1 and MN2, as well as a current source SC, which can be activated or deactivated by a control signal EN.

 More specifically, the respective gates of transistor MN1 and MN2 are connected to the two outputs of the operational amplifier AOP, while the respective drains of these transistors are connected to the two inputs of amplifier AOP. The sources of the two transistors are connected together to the output of the current source SC.

 It is assumed here that the order of the filter is greater than or equal to 3, for example equal to 4. Consequently, the diagram representing the monotonic evolution of the phase of the filter as a function of the frequency, presents a frequency point for which the low-pass filter phase goes through-180). Consequently, this filter is capable of oscillating when it is for example looped back on itself with a phase inversion.

 Although this loopback with phase inversion can be carried out directly between the output and the input of the operational amplifier, it is preferable to adopt the circuit illustrated in FIG. 2, in which the differential pair of MOS transistor provides a phase inversion.

 Furthermore, the input of the filter, that is to say in this case the differential input of operational amplifier AOP being a virtual ground, that is to say having a very high input impedance, this input can be controlled with the current delivered by the current source SC, and this without modifying the characteristic of the filter. The value of the current delivered by the current source SC, makes it possible to determine the amplitude of the oscillation of the filter. This current value will be chosen so that the filter works in the linear operating zone.

 The oscillation will be maintained by injecting the current delivered by the current source SC, alternately in the two branches of the differential pair of the MOS transistors.

 The invention uses here, in particular, the fact that a filter of order n phase of n? i: / 4 at the cutoff frequency. In other words, when

<Desc / Clms Page number 8>

 the filter has an order equal to 4, for example, it phase +/- TE at the cutoff frequency. In other words, the oscillation frequency of the filter is then theoretically its cut-off frequency.

 The filtering system according to the invention then comprises measurement means MM intended to determine the oscillation frequency of the filter, and correction means MCR able to correct the characteristics of the filtering device, that is to say say, in this case, the capacitive value of the capacitive network RC, taking into account the determined oscillation frequency and a pre-established relationship between the oscillation frequency and the theoretical cut-off frequency, so as to give the filtering device a cutoff frequency equal to the theoretical cutoff frequency to within a tolerance.

 An embodiment of these measurement means MM and of these correction means MCR is illustrated diagrammatically in FIG. 3.

 Thus, it can be seen that these measurement means MM include, for example, a CMP comparator header connected to the two outputs of the operational amplifier AOP. This comparator CMP performs a formatting of the output signal and delivers a square signal SCA.

 First processing means MT1 receiving, on the one hand, the signal SCA and, on the other hand, a reference clock signal CLKR, will count the number of periods of the reference clock signal CLKR between two rising edges of the SCA signal. This will determine the frequency of oscillation.

 In practice, the output signal delivered by the processing means MT1, which can for example be produced by means of a counter and logic gates, is a digital word MN3, for example on 9 bits, representative of the ratio between the frequency oscillation and frequency of the reference clock signal.

 The correction means MCR comprise, for example, a memory storing a preset correction table TAB defining, taking into account the order of the filter and its theoretical cut-off frequency, the correction to be made for each word MN3 received at the input of second MT2 processing means.

<Desc / Clms Page number 9>

 In practice, the second processing means comprise logical addressing means which will address the memory TAB with each word MN3 received in order to extract therefrom a correction word SCR on N bits (4, for example).

 This word SCR will then be delivered to the resistive-capacitive network RC to switch or not a certain number of capacitors of this network RC, so as to modify or not the capacitive value of this network RC according to the correction word SCR.

 The oscillation frequency of the filter will thus be able to be adjusted, so as to approach the theoretical cutoff frequency to within a tolerance.

 When this calibration phase is finished, control means, which can for example be incorporated within the baseband processor PR, then deactivate the current source SC by means of the signal EN. As a result, the filtering device no longer oscillates and performs its filtering function.

 In this regard, it should be noted that when the current source SC is deactivated, the operational amplifier AOP remains looped back via the gate / drain capacities of the MOS transistors MN1 and MN2. Therefore, and in order not to disturb the filtering function by this loopback, advantageously choose small MOS transistors so as to minimize the value of the gate / drain capacities.

 Those skilled in the art will be able to size the MOS transistors MN1 and MN2 in this regard, so as not to disturb the filtering device in its filtering function by a possible looping back of the filtering signal via the grid / drain capacities.

 As a variant, it would be possible to add to this embodiment a stage of the cascode type between the MOS transistors MN1 and MN2 and the inputs of the operational amplifier AOP, so as to achieve in the working phase of the filter, an isolation with respect to the grid / drain capacities of the MOS transistors MN1 and MN2.

 In practice, when the filtering system is incorporated in a cellular mobile phone for example, the calibration phase

<Desc / Clms Page number 10>

 will advantageously take place when the cellular mobile telephone is switched on, and possibly subsequently in response to a control signal delivered by the baseband processor PR.

 The invention is not limited to the embodiments and implementation which have just been described, but embraces all variants thereof.

 Thus, if the order of the filter is equal to 4 or is a multiple of 4, the filter oscillates at its cutoff frequency. This is not the case when the order of the filter, while being greater than or equal to 3, is different from 4 or a multiple of 4. In this case, the filter oscillates at a frequency different from the frequency of cutoff, and which depends on the quality coefficient of the filter. This quality coefficient being known, it should then be taken into account in the correction table TAB contained in the memory of the correction means.

 If the order of the filter is less than 3, for example equal to 2, there will then be available for the calibration phase another filter identical to the filter to be calibrated, in series with this filter to be calibrated.

 In practice, in the case of a cellular mobile telephone for example, since there are two processing channels 1 and Q, it will be advantageous to connect in series, in the calibration phase, the two homologous filters respectively located on the two processing channels 1 and Q, for example filters FF10 and FF20. This set of two filters will then be looped back on itself via the looping stage ETRB for example, so as to oscillate. The two filters will then be calibrated simultaneously.

 Furthermore, in order to facilitate the start of the oscillation, provision may be made to temporarily connect to one of the drains of the transistors MN1 and MN2, an additional current source injecting at start-up an additional current into one of the branches of the differential pair, so as to unbalance the system and thus favor the start of the oscillation.

 The invention has also been described here with a so-called low-pass filter. Of course, it can be applied to a high-pass type filter.

<Desc / Clms Page number 11>

 The invention can also be applied successively to a high-pass filter and a low-pass filter, connected in series and together forming a band-pass filter.

 Finally, although a filter with a resistive-capacitive network has been described here, the invention also applies to a filter having an inductive-capacitive network.

Claims (12)

1-Method for controlling the operation of a monotonic phase filtering device having a theoretical cutoff frequency, comprising a calibration phase in which the filtering device (FF10) is operated as an oscillator, the oscillation frequency is determined of the filtering device, and the characteristics of the filtering device are corrected taking into account the determined oscillation frequency and a predetermined relationship between the oscillation frequency and the theoretical cutoff frequency, so as to give the filtering device a cutoff frequency equal to the theoretical cutoff frequency to within a tolerance, and a work phase in which the filtering device performs its filtering function.  2-A method according to claim 1, characterized in that when the filter device (FF10) has an order greater than or equal to 3, the filter device is operated as an oscillator by looping it back on itself with an inversion of phase, and by the fact that when the filter device has an order of less than 3, an identical auxiliary filter device (FF20) is connected in series to the filter device, and the assembly formed by the two filter devices is operated as an oscillator by looping this assembly back on itself with phase inversion.  3-A method according to claim 2, characterized in that the input of the filtering device is a virtual mass, and by the fact that the loopback is piloted with a current injected onto said virtual mass so as to adjust the amplitude oscillations.  4-Method according to one of the preceding claims, characterized in that the cutoff frequency of the filtering device is defined by a resistive-capacitive network (RC), and by the fact that the characteristics of the filtering device are corrected by modifying the capacitive value of the resistive-capacitive network.  5-Electronic filtering system, characterized in that it comprises <Desc / Clms Page number 13>  - a monotonic phase filtering device (FF10) having a theoretical cut-off frequency, first controllable means (ETRB) connected to the filtering device and capable of being activated or deactivated in response to a control signal, so as to do operate the filtering device as an oscillator when activated, the filtering device performing its filtering function when the first means are deactivated, measuring means (MM) capable of determining the oscillation frequency of the filtering device, - correction means (MCR) capable of correcting the characteristics of the filtering device taking into account the determined oscillation frequency and a pre-established relationship between the oscillation frequency and the theoretical cut-off frequency, so as to give the device filtering a cutoff frequency equal to the theoretical cutoff frequency to within a tolerance, and - control means ( PR) capable of delivering the control signal (EN).  6-System according to claim 5, characterized in that the filtering device has an order greater than or equal to 3, and in that the first means comprise a loopback stage with phase inversion (ETRB) connected between the output and the input of the filtering device.  7-System according to claim 5, characterized in that the filtering device has an order less than 3, and in that the first means comprise a loopback stage with phase inversion (ETRB) connected between the output and the input of the filtering device, this looping stage comprising an auxiliary filtering device identical to said filtering device, the two filtering devices being connected in series.  8-System according to claim 6 or 7, characterized in that the input of the filtering device is a virtual mass, and by the fact that the looping stage comprises means (SC, MN1, MN2)) <Desc / Clms Page number 14>  able to control the loopback with a current injected onto said virtual mass so as to adjust the amplitude of the oscillations.  9-System according to claim 8, characterized in that the filtering device has a differential structure and comprises a differential operational amplifier (AOP), and by the fact that the control means comprise a differential pair of transistors (MN1, MN2 ), whose respective sources or transmitters are connected, whose respective grids or bases are connected to the two outputs of the differential amplifier, and whose respective drains or collectors are connected to the two inputs of the differential amplifier, and a source of current (SC) capable of being activated or deactivated in response to the control signal, and connected to the sources of the two transistors.  10-System according to one of claims 5 to 9, characterized in that the filtering device comprises a resistive-capacitive network (RC) defining the cut-off frequency, and by the fact that the correction means are able to correct the characteristics of the filtering device by modifying the capacitive value of the resistive-capacitive network.  11-System according to one of claims 5 to 10, characterized in that it is produced in the form of an integrated circuit.  12-Remote terminal of a wireless communication system, in particular cellular mobile telephone, characterized in that it incorporates a filtering system according to one of claims 5 to 11.